Use GSS occupational recodes to update estimates of intergenerational persistence.
Start with SEI vs. Hauser-Warren, same-sex parent to current occ
Rescale Hauser-Warren to ranks. Update estimates.
Outcomes are scewed, esp., at high and low ends of origin distribution (more for men than women). Switch to robust (median) regression.
After peer review, I dropped the rank approach and used the more conventional SEI approach, per reviewer suggestions (and objections to details of the ranking procedure).